Clusters of databases have become an attractive platform for large information systems. They offer high performance, 'scale-out' scalability, fault tolerance, and a superior cost/performance ratio. The challenge however is to build them in a way that all these properties are present at the same time, in particular if used for on-line analytical processing. This book investigates central architectural issues and performance aspects of clusters of databases for usage in a decision support scenario. It presents a scalable, middleware-based cluster architecture and develops innovative algorithms for high-performance query routing based on approximated cache states. Another important contribution is a novel approach to coordinated replication management for large clusters. The combination of these techniques allows for efficient on-line analytical processing with a cluster of databases where clients can trade result freshness for query performance while even being capable of analyzing up-to-date data.
